Related papers: MultiPath Memory Access: Breaking Host-GPU Bandwid…
Multi-accelerator servers are increasingly being deployed in shared multi-tenant environments (such as in cloud data centers) in order to meet the demands of large-scale compute-intensive workloads. In addition, these accelerators are…
Training large language models (LLMs) in the cloud faces growing memory bottlenecks due to the limited capacity and high cost of GPUs. While GPU memory offloading to CPU and NVMe has made large-scale training more feasible, existing…
In this paper, we propose bit-interleaved multiple access (BIMA) to enable Internet-of-Things (IoT) networks where a massive connection is required with limited resource blocks. First, by providing a true power allocation (PA) constraint…
The two main thrusts of computational science are more accurate predictions and faster calculations; to this end, the zeitgeist in molecular dynamics (MD) simulations is pursuing machine learned and data driven interatomic models, e.g.…
The proliferation of camera-enabled devices and large video repositories has led to a diverse set of video analytics applications. These applications rely on video pipelines, represented as DAGs of operations, to transform videos, process…
Memory tiering provides a cost-effective solution to increase memory capacity, utilization, and even bandwidth. Memory tiering relies on system software for memory profiling, detection of frequently accessed pages, and page migration. Such…
Memory and logic integration on the same chip is becoming increasingly cost effective, creating the opportunity to offload data-intensive functionality to processing units placed inside memory chips. The introduction of memory-side…
It is often said that one of the biggest limitations on computer performance is memory bandwidth (i.e."the memory wall problem"). In this position paper, I argue that if historical trends in computing evolution (where growth in available…
Distributed ML workloads rely heavily on collective communication across multi-GPU, multi-node systems. Emerging scale-up fabrics, such as NVLink and UALink, enable direct memory access across nodes but introduce a critical destination-side…
CUDA Unified Memory improves the GPU programmability and also enables GPU memory oversubscription. Recently, two advanced memory features, memory advises and asynchronous prefetch, have been introduced. In this work, we evaluate the new…
PIM architectures aim to reduce data transfer costs between processors and memory by integrating processing units within memory layers. Prior PIM architectures have shown potential to improve energy efficiency and performance. However, such…
This work introduces, for the first time, non-orthogonal multiple access (NOMA) into short-packet communications to achieve low latency in wireless networks. Specifically, we address the optimization of transmission rates and power…
Future 5G communication systems require more demanding performances than the existing cellular communication systems, e.g., 10 to 100 Mbps user data rate and much larger cellular spectrum efficiency. The well-used multiple access methods…
Various verticals in 5G and beyond (B5G) networks require very stringent latency guarantees, while at the same time envisioning massive connectivity. As a result, choosing the optimal multiple access (MA) technique to achieve low latency is…
End-to-end Vision-Language-Action (VLA) models for autonomous driving unify perception, reasoning, and control in a single neural network, achieving strong driving performance but requiring 20-60GB of GPU memory-far exceeding the 12-16GB…
The Lattice Boltzmann method (LBM) for solving fluid flow is naturally well suited to an efficient implementation for massively parallel computing, due to the prevalence of local operations in the algorithm. This paper presents and analyses…
Cutting-edge embedded system applications, such as self-driving cars and unmanned drone software, are reliant on integrated CPU/GPU platforms for their DNNs-driven workload, such as perception and other highly parallel components. In this…
The torrential influx of floating-point data from domains like IoT and HPC necessitates high-performance lossless compression to mitigate storage costs while preserving absolute data fidelity. Leveraging GPU parallelism for this task…
Many performance critical systems today must rely on performance enhancements, such as multi-port memories, to keep up with the increasing demand of memory-access capacity. However, the large area footprints and complexity of existing…
We present the first comprehensive study of latent multi-head attention (MLA) for small language models, revealing interesting efficiency-quality trade-offs. Training 30M-parameter GPT models on 100,000 synthetic stories, we benchmark three…